Several companies working on 5G networks and services — Fujitsu, Google Cloud, Mavenir, Radisys and Wind River — are teaming up with semiconductor company Nvidia to create solutions for Nvidia’s AI-on-5G platform to help speed the development of a variety of use cases, Nvidia announced today.
While no mobile network operators were listed among this group, Nvidia said it is working closely with operators, and that the AI-on-5G solutions that result will directly benefit carriers. For example, Fujitsu said as part of Nvidia’s announcement that it will leverage the AI-on-5G platform to develop open RAN solutions for operators. In particular, Fujitsu and Nvidia will be helping NTT DOCOMO move toward open RAN in 5G.
“Today, MNOs are already using Nvidia EdgeAI solutions to provide smart cities, cloudXR, cloud gaming and other services,” said Ronnie Vasishta, senior vice president of Telecom at Nvidia, in an email. “The next step for operators is virtualized 5G connectivity, and they’re evaluating Nvidia GPUs and DPUs for 5G RAN and core to implement a converged AI-on-5G platform. In terms of names that we’re working with on AI applications, we’ve publicly announced NTT East, NTT DOCOMO, Vodafone and British Telecom.”
Vasishta also noted that Nvidia is engaged closely enough with Verizon Media and Vodafone that both of those operators are speaking on panels during Nvidia’s GTC 2021 virtual event this week.
Jack Gold, founder and principal analyst at J. Gold & Associates, told FierceWireless, “All the carriers will be using AI with 5G. AI will help them model their networks to look for anomalies, help them service and sell to customers and also with optimizing network slicing. Nvidia is making it easier for them to do all of that.”
The AI-on-5G platform uses Nvidia’s Aerial software development kit and the company’s BlueField-2 A100 — a converged card that combines graphic processing units and data processing units, including Nvidia’s “5T for 5G” solution. The resulting converged platform can handle both 5G and edge AI computing for high-performance 5G RAN and AI applications, while reducing development time and speeding delivery of AI applications to enterprises, according to Nvidia.
Mavenir, Radisys, Wind River
Among the other co-development efforts announced, Mavenir is building two 5G vRAN systems based on the Aerial SDK, and it will target both 5G network operators and enterprise AI opportunities in private 5G networks with what Mavenir and Nvidia described as “a hyperconverged enterprise 5G solution.”
Meanwhile, Radisys and Wind River plan to deliver Nvidia Aerial AI-on-5G solutions for enterprise 5G and industrial 5G networks, while Google Cloud is working with Nvidia to extend Google’s Anthos application management platform to the network edge to help customers speed the deployment of enterprise-grade, containerized applications.
“Operators will be able to benefit in many ways from implementing Nvidia’s AI-on-5G into the network,” Vasishta said in the email. “From the perspective of hyperconverged 5G, you’ll see a convergence of functions into space and power constrained environments – 5G core, 5G vRAN, DPI, vFirewall, front haul termination, timing sync, AI-based MEC workloads. Then there’s AI for network management and optimization… But the key here is that AI is the new over-the-top killer app - much like streaming was to 4G.”
Nvidia's announcements this week come as the company continues to work toward its planned acquisition of Arm, a deal that has faced intense regulatory scrutiny, as well as questions about how it might affect semiconductor competition and innovation on mobile and other sectors.